On Mon, 2013-03-04 at 14:15 -0500, Vivek Goyal wrote:
> I am just brain storming and throwing some ideas and see if soemthing
> makes sense. I agree that allowing one policy only makes it very
> restrictive (while simplifying the implementation).
Agreed, lets try again ... I think we are actually getting closer.
Without the memory locking or caching issues, would you agree that both
the builtin 'secureboot' and the 'ima_appraise_tcb' policies meet the
secure boot needs? If both policies are acceptable, then we could
define the builtin 'secureboot' policy as the default policy, which
could be replaced with the 'ima_appraise_tcb' policy, if specified on
the boot command line. This would eliminate any need for merging of
rules or rule flags.
To address the memory locking and caching issues, we could define a new
extended attribute type called IMA_XATTR_SECURE_BOOT.
enum evm_ima_xattr_type {
IMA_XATTR_DIGEST = 0x01,
EVM_XATTR_HMAC,
EVM_IMA_XATTR_DIGSIG,
};
and set a corresponding flag in iint->flags. The flag could then be the
bases for setting up any special secureboot requirements, like memory
locking and no caching.
diff --git a/security/integrity/ima/ima_appraise.c
b/security/integrity/ima/ima_appraise.c
index ddeadc7..6ec1575 100644
--- a/security/integrity/ima/ima_appraise.c
+++ b/security/integrity/ima/ima_appraise.c
@@ -172,6 +172,8 @@ int ima_appraise_measurement(int func, struct
integrity_iint_cache *iint,
}
status = INTEGRITY_PASS;
break;
+ case IMA_XATTR_SECURE_BOOT:
+ iint->flags |= IMA_SB;
case EVM_IMA_XATTR_DIGSIG:
iint->flags |= IMA_DIGSIG;
rc = integrity_digsig_verify(INTEGRITY_KEYRING_IMA,
As originally intended, the policy defines which files are appraised,
not how they are appraised. The extended attribute defines how the file
is to be appraised.
thanks,
Mimi
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
